Skip to content
This repository has been archived by the owner on Sep 15, 2022. It is now read-only.

canonical/elasticsearch-k8s-operator

Repository files navigation

elasticsearch-operator

Description

The Elasticsearch Operator provides a distributed search and analysis solution using Elasticsearch.

Setup

Increase the maximum number of virtual memory areas on your host system. On a Linux system this can be done using the command line

sudo sysctl -w vm.max_map_count=262144

For a more permanent change edit /etc/sysctl.conf.

Install Dependencies and Build

To build the charm, first install charmcraft, juju and microk8s

snap install charmcraft
snap install juju --classic
snap install microk8s --classic 

Then in this git repository run the command

charmcraft build

Usage

juju deploy ./elasticsearch.charm

To scale up:

juju add-unit -n 2 elasticsearch

Note: When the total number of nodes in the cluster is 2, split brain is possible. If there are currently two nodes, be sure to use juju add-unit to scale up to a functional HA cluster. For more information about how Elasticsearch handles quorum and the effects of split brain, take a look at the Elasticsearch docs.

To check the status of the cluster:

# 1. wait until all units and application is active
watch -c juju status --color
# 2. copy the application IP address from the juju status output
# 3. check the health of the cluster
curl -X GET http://{APP_IP}:9200/_cat/health?v&pretty

Developing

Use your Python 3 development environment or create and activate a virtualenv, and install the development requirements,

virtualenv -p python3 venv
source venv/bin/activate
pip install -r requirements-dev.txt

Testing

Just run run_tests:

./run_tests